Of the 4 million infants born in the United States each year, about 20,000 (5 out of 1,000) die during the neonatal period and another 10,000 die after the neonatal period but before their first birthday. Some of these deaths are caused by conditions that can be treated successfully if they are diagnosed promptly. One factor that may influence whether such conditions are diagnosed quickly is the length of a well-appearing newborn's postpartum stay, because signs of congenital heart disease, infection, and other health problems may not be evident until two or more days after delivery. Although many studies have examined health effects of early postpartum discharge, few studies have considered mortality as an outcome. The investigators propose to assess the association between early discharge and neonatal mortality using an administrative database from California that captures about three million births between 1991 and 1999. The primary objective of the proposed analysis is to test the hypothesis that newborns discharged early are at increased risk of dying during the neonatal period relative to newborns with longer stays. The investigators will perform unadjusted analyses, multivariate analyses to control for observed confounders, and instrumental variable estimation to mitigate biases due to unobserved confounders. The investigators will also perform analyses of high-risk sub-populations. The secondary objective of the proposed analysis is to test the hypothesis that newborns discharged early are at increased risk of hospital readmissions during the neonatal period relative to newborns with longer stays. The investigators will also describe time trends in early discharge, neonatal mortality, and hospital readmission rates. An improved understanding of these issues will provide policymakers, health plans, hospitals, providers, and patients with information that can be used to help make more informed decisions about newborn lengths of stay.